【作者】 韩璐；

【导师】 黄岚；

【作者基本信息】 吉林大学 ， 软件工程（专业学位）， 2015， 硕士

【摘要】 在信息爆炸时代，信息量急剧上升，如何在浩如烟海数据中找到人们需要的信息成为一个热门的研究课题。推荐系统提供了信息过滤技术，很好的帮助用户在面对大量信息时能够节约时间成本，并且有效的提高用户的满意度，因此本文针对电商平台设计了个性化推荐系统。本文从实际应用出发，重点介绍根据用户的兴趣偏好利用个性化推荐技术，向用户推荐用户可能感兴趣的信息，从而根据不同用户进行个性化服务。个性化推荐技术的应用领域非常广泛，比如电子商务领域，通过个性化推荐系统可以提高电子商务网站的购买率。个性化推荐系统技术从本质上来说就是一种信息过滤技术，它是一种融合了各种数据挖掘算法并结合用户相关信息，完成符合用户兴趣或潜在兴趣的综合系统推荐技术。根据推荐算法的不同，推荐系统分为不同的类型。在传统的推荐系统分类中，推荐系统分为基于内容的推荐系统，协同过滤推荐系统，混合推荐系统。由于每种推荐算法的使用范围不同，使用条件不同，导致对于同样的信息推荐，各个推荐算法的推荐效果是不同的。在实际推荐系统应用中，一般都趋向于混合推荐系统，即将各种推荐算法通过混合的方法将各个算法的长处融合到实际的推荐过程中，有效地提高了推荐质量。本文研究内容是帮助用户在海量商品信息数据中自动的获取用户感兴趣的商品，从而避免用户受到过多其他信息的干扰。首先介绍了相关研究背景和国内外的研究现状，进而给出本文要研究与实现的主要工作。其次，针对本文要实现的部分，采用适当的开发技术，保证本文系统开发能够顺利进行。经过需求分析和系统定位，本文在传统算法的基础上提出了改进算法，在介绍了总体设计与数据库表单设计之后，文中最后通过系统界面效果展示了系统的成果。其中，本文的主要工作包含对个性化推荐系统的算法的研究，结合实际应用需求，通过研究结果与需求分析相结合，设计了一个能有效为用户提供个性化推荐的推荐管理系统，并对系统进行实现。更多还原

【Abstract】 With the developing of internet technology, more and more people dependent onaccess to information by the internet. In the era of information explosion, because of arise in the amount of information, how to find out useful information from massivedata become a hot research topic. Recommended system provides informationfiltering technology, which can help users to save time in the face of large amounts ofinformation, and it can improve customer satisfaction effectively, this paperproposes a personalized recommendation system.Personalized recommendation technology recommend information to userswhich they may be interested in based on their interest or preference, so it can providepersonalized service according to different users. Personalized recommendationtechnology broad range of applications, such as e-commerce, e-commerce sites canincrease purchase through personalized recommendation system. Personalizedrecommendation system technology is essentially an information filtering technology,it is an integrated system with a fusion of a variety of data mining algorithmscombined with user-related information, to predict to interest or potential interest ofusers. Recommended system is divided into different types depending on therecommendation algorithm. In the traditional recommendation system classification,it is divided into content-based systems, collaborative filtering system, hybridsystem. Because of the different algorithms are used in different conditions, therecommend effect is not the same one though the recommendation algorithm are usedsame information.Generally, in the practical application of recommendation systems,it tend to hybrid recommendation system, which is a method of mixing variousrecommendation algorithm, it can improve the recommendation result effectively. The content of paper is to research how to help a user to access the interestedinformation automatically in the mass commodity product information data, whichcan avoid interference. Firstly, this paper introduces the research background andresearch status, then it gives the main work of this paper. Secondly, part of the articleto be achieved, we use the appropriate technology to ensure the conduct of systemdevelopment smoothly. After positioning requirements analysis and systems, it isproposed to improve the algorithm on the basis of the traditional algorithm, the papershows the final results by the system interface after the introduction of the databasedesign and form design. The main work of this paper contains a study of personalizedrecommendation system algorithm, combined with practical application requirements,it is designed to provide users with an effective personalized recommendationsmanagement system through a combination of research results and needs analysis,now the system is realized.更多还原